shevalier wrote on 2026-08-08, 11:24:

Near the RAM slots, there is an unpopulated footprint for an electrolytic capacitor. If it's on the +3.3V rail, I would try runn […]
Show full quote

Near the RAM slots, there is an unpopulated footprint for an electrolytic capacitor.
If it's on the +3.3V rail, I would try running a jumper wire to it.
That way the run will be much shorter, and you can use a thinner wire.
Plus, it will look much cleaner.

But again, I don't have this board.
I don't know where this pad goes.
This needs to be checked with a multimeter.

Shevalier if you mean this one:

^ It's a CPU VRM output spot that they just left unpopulated (like the one to its extreme left in this pic, that I've circled in blue). They both are in parallel with the six that came OEM populared.

zami555 wrote on 2026-08-08, 11:32:

TELVM wrote on 2026-08-01, 08:44:

^ I bet that "brilliant" idea came from the same "genius" that made the "exquisite" selection of capacitors. 👎

I bet the decision was made by management and financial departments of Gigabyte. Engineers usually try to do their best with designing something... And then others come in and decide. It's just a few words from me to support the engineer who worked on this design 😉

^ You're completely right, the penny-pinching is always imposed by the bean-counters at the top.

I'm sure back in the day Gigabyte engineers could, and wanted to, do miles better.

.

BTW, if anything I'm a Gigabyte fan 😀 . I'm writing this from a X570 Aorus Master, and also own several other Gigabyte boards and GPUs. 👍

TELVM wrote on 2026-08-08, 11:47:

Shevalier if you mean this one: […]
Show full quote

shevalier wrote on 2026-08-08, 11:24:

Near the RAM slots, there is an unpopulated footprint for an electrolytic capacitor. If it's on the +3.3V rail, I would try runn […]
Show full quote

Near the RAM slots, there is an unpopulated footprint for an electrolytic capacitor.
If it's on the +3.3V rail, I would try running a jumper wire to it.
That way the run will be much shorter, and you can use a thinner wire.
Plus, it will look much cleaner.

But again, I don't have this board.
I don't know where this pad goes.
This needs to be checked with a multimeter.

Shevalier if you mean this one:

^ It's a CPU VRM output spot that they just left unpopulated (like the one to its extreme left in this pic, that I've circled in blue). They both are in parallel with the six that came OEM populared.

And the next two near the chipset?

shevalier wrote on 2026-08-08, 12:21:

And the next two near the chipset?

They're in parallel with each other and connect to pins in all three memory slots. They both are also in parallel with the two caps right north from the LX8383 (which I believe feeds ~3.3V juice to all of them):

While we're at it.

Keep in mind that I'm not a pro, just a hobbyst, and I may well be wrong.

But to the best of my knowledge these are the voltages handled by every cap on the GA-5AX Rev. 4.1 :

BTW right now I'm recapping this board and have it lying on the operating table, so if anyone thinks of anything else worth scrutinising, now is the moment 😀 .

^ The mod works fine. 👍

Voltage measured at the location now vacated by the LX8383 linear voltage regulator reads 3.35V .

As expected, the +5V rail has received a slight boost in voltage from the unloading.

And overall system power draw from the wall has decreased by a bunch of watts. 😀

If you ever decide to move it to the front, it's relatively trivial to pull the plastic ATX connector shroud off.
Sneak the wire in, trim the shroud to fit.

In the interest of advancing mankind's knowledge, I've perpetrated some mad scientist experiments.

After this direct +3.3V wire mod, this GA-5AX is able to POST with every specimen in my highly exclusive collection of premium, rare and valuable AGP graphics cards, i.e.:

S3 Trio 3D/2X

Nvidia TNT2 M64

ATi Radeon 9200 SE 64-bit

Nvidia GeForce 6200A 64-bit

More seriously now, to my surprise it even manages to POST with a relatively power-hungry Radeon 9800 Pro 128MB (with Molex connector plugged in the card for extra +5V & +12V external power from PSU, of course).

I'm not exploring this combination beyond POST because, to put it mildly, "it's a considerable mismatch" 😀 . But POST it does.

.

Interestingly, with this same 9800 Pro my Tualatinized Slot-1 440BX board is unable to POST (continuous one long beep, two short beeps = Video card initialisation problem).

And now I think I have the culprit identified:

Guess who is gonna receive a direct +3.3V wire mod next ...

TELVM wrote on 2026-08-22, 09:39:

with Molex connector plugged in the card for extra +5V & +12V external power from PSU, of course

not.
When people see an external Molex connector on a graphics card, their brains just completely shut down.
THEY DIDN'T DO IT LIKE THAT BACK THEN, this isn't an RTX 5090.The Molex connector only provides a portion of the power; the rest (+3.3V) is still drawn from the AGP slot.

Interestingly, with this same 9800 Pro my Tualatinized Slot-1 440BX board is unable to POST (continuous one long beep, two short beeps = Video card initialisation problem).

An absolutely extraordinarily lucky sample of a Radeon 98(7)00 will post on a 440BX at 133MHz FSB only if you have an absolutely magical sample (specifically the sample, not the model) of the motherboard.The Radeon 9800 cannot handle an 89MHz AGP frequency with 3.3V signaling lines (not AGP 2x 1.5V — though that's also an option for AGP 4x systems).The 9100/9200 runs without issues, the 9600 won't even fit in the slot, and the 9800 simply doesn't work.Which is exactly why all 440BX owners are suffering.

shevalier wrote on 2026-08-22, 10:23:

... The Molex connector only provides a portion of the power; the rest (+3.3V) is still drawn from the AGP slot ...

I agree completely. When have I said otherwise? 😀

shevalier wrote on 2026-08-22, 10:23:

... The Radeon 9800 cannot handle an 89MHz AGP frequency with 3.3V signaling lines ...

I know. Please hold your horses my friend. 😀

My apologies, I was a bit too categorical.
There's always room for a miracle, and if your Radeon 9800 runs stably on a 133 MHz bus with the 440BX chipset, I’ll be incredibly jealous. 😀

No need to apologize my friend. You're in a combat zone, where no-nonsense approach and getting straight to the point goes with the territory.

I thank you for all your help, I'm learning a lot from your posts. 👍